LTO 8 mounting orientation effects on Tape drive performance

The topic of the preferred mounting orientation of tape drives has been discussed for decades, clearly the drive mounting orientation will alter the drive head mass effect on the servo loop performance when the head is suspended by the actuator or resting on the guides. Also the drive mounting orientation could affect the dynamics of the tape lateral tape motion as it spools onto the reels and even the head to tape interface could affect the sensitive read/write performance.

Due to the complexity of the system interaction, the first order of this analysis was to design a simulated customer test to determine if we can detect any orientation effects. Over the years, number of tests were conducted on various DLT drives and even decades ago, the test data showed a preferred drive mounting orientation, considering the LTO8 has similar drive architecture to DLT, the question regarding the effect of tape drive mounting orientation on the drive performance of LTO drives has remained largely unanswered.

This study was recently commissioned by Magstor to quantify the effect of the mounting orientation on the LTO8 drive performance. The tests were conducted on a new IBM LTO8 Tape Drive supplied by Magstor, (Magstor PN SAS-HL8-8088), using the standard Tape drive Doctor advanced test routines.

First few words about Tape Drive Doctor tool. TDD was originally developed and patented by Quantum Corporation and later advanced by Saliba Technologies Corporation. The objective of this tool is to quickly test LTO drives performance during simulate customer read/write operations and cartridge read/write interchange using calibrated cartridges. This tools has been used extensively by drive repair companies and more information regarding the technology and patents can be found at TapedriveDoctor.com.

TDD basic routine consist of reading and writing customer data at specific tape locations, by measuring the time of various operations, a figure of merit of the drive performance is reported. TDD also interrogate the drive error rates and servo defects so at the end of the test, we have a full 360 degrees results representing tape drive performance. All TDD test cartridges are calibrated to a gold standard, and the results are purely the drive performance. The quality value is linearized and normalize. The value of 3, represent the centerline drive, a lower value is better than average and higher value represent a worse than average drive. For complete picture, we will report the raw data reported by the drive, the actual process time and normalize performance.

For the test, a new IBM LTO8 drive serial number YY1097006353 supplied by MagStor was used. The two orientations tested were according to the instruction of the drive manufacture, first the horizontal direction two passes. Second, two passes in the vertical direction. Special attention was taken to insulate the drive from any possible outside vibration to purely measure the effect of the mountain orientation.

For the first set of tests, the drive mounted in the horizontal direction.

Below are the timing data as reported by TDD test portion of writing 5000 blocks repeated 8 cycles when the drive mounted in the horizontal direction.

LTO8 First test horizontal reverse direction

01/18/2021 15.38.53.328 Wr Rev Process Time Cycle 1 in Milliseconds
01/18/2021 15.38.53.375 2 2 2 2 2 2 2 2
01/18/2021 15.38.53.406 gCPW_TimeRev Normalized
01/18/2021 15.38.53.468 3.022 3.000 3.000 3.000 3.000 3.000 3.000 3.000

LTO8 Second test horizontal reverse direction

01/18/2021 16.42.56.859 2 2 2 2 2 2 2 2
01/18/2021 16.42.56.890 gCPW_TimeRev Normalized
01/18/2021 16.42.56.953 3.000 3.000 3.000 3.000 3.000 3.000 3.000 3.000

LTO8 First test horizontal forward direction

01/18/2021 15.40.28.796 2 2 2 2 2 2 2 2
01/18/2021 15.40.28.828 gCPW_TimeFwd Normalized
01/18/2021 15.40.28.890 3.000 3.000 3.000 3.000 3.000 3.000 3.000 3.000

LTO8 Second test horizontal forward direction

01/18/2021 16.44.32.375 2 2 2 2 2 2 2 2
01/18/2021 16.44.32.406 gCPW_TimeFwd Normalized
01/18/2021 16.44.32.468 3.000 3.000 3.000 3.000 3.000 3.000 3.000 3.000

For the second set of tests, the drive mounted in the vertical direction.
Below are the timing data as reported by TDD test portion of writing 5000 blocks repeated 8 cycles when the drive mounted in the vertical direction.

LTO8 First test vertical writing in the reverse direction

01/18/2021 15.51.41.500 8 3 2 2 2 2 2 2
01/18/2021 15.51.41.531 gCPW_TimeRev Normalized
01/18/2021 15.51.41.609 12.176 4.773 3.000 3.000 3.000 3.000 3.000 3.000

LTO8 Second test vertical writing in the reverse direction

01/18/2021 16.09.25.796 7 2 2 2 2 2 2 2
01/18/2021 16.09.25.812 gCPW_TimeRev Normalized
01/18/2021 16.09.25.875 11.140 3.000 3.000 3.000 3.000 3.000
3.000 3.000

LTO8 First test vertical writing in the forward direction

01/18/2021 15.53.39.531 2 2 2 2 2 2 2 2
01/18/2021 15.53.39.578 gCPW_TimeFwd Normalized
01/18/2021 15.53.39.656 3.055 3.000 3.000 3.000 3.000 3.000
3.000 3.000

LTO8 second test vertical writing in the forward direction

01/18/2021 16.21.44.125 2 2 2 2 2 2 2 2
01/18/2021 16.21.44.171 gCPW_TimeFwd Normalized
01/18/2021 16.21.44.234 3.000 3.000 3.000 3.000 3.000 3.000
3.000 3.000

Horizontal, first pass LTO8 servo data

01/18/2021 15.39.52.031 Servo stop write Fwd : 0
01/18/2021 15.39.52.046 Servo stop write Rev : 0
01/18/2021 15.39.52.062 Write skip Fwd : 363
01/18/2021 15.39.52.078 Write skip Rev : 1966

Horizontals, second pass LTO8 servo data

01/18/2021 16.21.07.437 Servo stop write Fwd : 0
01/18/2021 16.21.07.453 Servo stop write Rev : 15
01/18/2021 16.21.07.468 Write skip Fwd : 370
01/18/2021 16.21.07.484 Write skip Rev : 1964

Vertical, first pass LTO8 servo data

01/18/2021 15.53.02.812 Servo stop write Fwd : 0
01/18/2021 15.53.02.828 Servo stop write Rev : 152
01/18/2021 15.53.02.859 Write skip Fwd : 427
01/18/2021 15.53.02.875 Write skip Rev : 2363

Vertical, second pass LTO8 servo data

01/18/2021 16.10.41.390 Servo stop write Fwd : 0
01/18/2021 16.10.41.406 Servo stop write Rev : 167
01/18/2021 16.10.41.421 Write skip Fwd : 404

Tape drive doctor detected significant 300% slower timing write performance degradation in the reverse direction when the drive was mounted in the vertical direction as opposed to the horizontal direction. Upon further analysis of the dive raw data, the cause was determined to be the servo loop optimization.

01/18/2021 16.10.41.421 Write skip Rev : 2377

Based on the data from testing the new IBM LTO8 drive, when the drive is mounted in the vertical direction, the drive delivered significantly worse performance than in the horizontal position. In this calibrated test, LTO Tape Driver Doctor, during repeated tests, measured as much as 3X slower performance when the drive was mounted in the vertical direction as compared the horizontal direction. Furthermore, the drive reported an order of magnitude increase in write servo skips in this drive. Large servo skip count indicates the potential for large losses of cartridge capacity that will reduce the total useable capacity of the LTO8 tape drive. Based on this test, we recommend mounting IBM LTO8 drives in the horizontal direction.